Regulation of Axl receptor tyrosine kinase expression by miR ...

ORIGINAL ARTICLE

Regulation of Axl receptor tyrosine kinase expression by miR-34a and

miR-199a/b in solid cancer

G Mudduluru1, P Ceppi1,2, R Kumarswamy1, GV Scagliotti2, M Papotti2 and H Allgayer1

1Department of Experimental Surgery and Molecular Oncology of Solid Tumors, Medical Faculty Mannheim, University ofHeidelberg, and German Cancer Research Center (DKFZ)-Heidelberg, Mannheim, Germany and 2Department of Clinicaland Biological Sciences, University of Turin at San Luigi Hospital, Orbassano, Italy

Axl is a receptor that induces proliferation, migration andinvasion in cancer. In this study, we show that specificmicroRNAs (miRNAs) target the 30-UTR of Axl.Luciferase-reporter assays with wild-type and deletedmiR-34 and miR-199a/b seed sequences of Axl 30-UTRconfirmed the specificity of targeting. An inverse correla-tion between Axl protein and miR-34a expression in apanel of non-small cell lung cancer (NSCLC), colorectalcancer (CRC) and breast cancer (BRC) cell lines wasobserved, while miR-199a/b expression was completelysuppressed. Pre-miR transfection inhibited in vitro migra-tion and invasion and, in vivo, reduced the number ofdistant lung- or liver-metastases in a chorion-allantoic-membrane (CAM) assay. Moreover, methylation-specificPCR on bisulfite-converted DNA obtained from the celllines showed that the miR-34a promoter methylationstatus was inversely correlated with its expression, andthat miR-199a/b promoter regions were methylated in allcells tested. In a panel of NSCLC tissues (n¼ 44), miR-34a and miR-199a/b were found to be downregulated andsignificantly co-expressed. A lower expression of all threemiRs was significantly associated with squamous histo-types, and, in a preliminary series, NSCLC patients withmiR-34a upregulation showed a positive associationtowards a longer survival. These results indicate thatAxl receptor expression can be regulated by miR-34a andmiR-199a/b, which are suppressed by promoter methyla-tion in solid cancer cells.Oncogene (2011) 30, 2888–2899; doi:10.1038/onc.2011.13;published online 14 February 2011

Keywords: non-small cell lung cancer; breast cancer;colorectal cancer; Axl; microRNA

Introduction

Cancer is a complex disease occurring as a result ofprogressive accumulation of genetic aberrations and

epigenetic changes that enable escape from normalcellular and environmental controls (Weinberg, 1995).Ninety percent of patient deaths in solid tumors are dueto metastasis, therefore, an understanding of the processand pathogenesis at the systemic, cellular and molecularlevels is one of the most ambitious goals in cancerresearch (Gupta and Massague, 2006; Esteller, 2007).Epigenetic changes in the promoters of classical andsmall non-coding (miRNA) genes has a pivotal role inthe acquisition of tumorigenic and metastatic properties(Feinberg and Tycko, 2004; Dumont et al., 2008;Lujambio et al., 2008). Methylation occurs in specificgenomic areas called CpG islands, especially within thepromoter region of a gene, thereby preventing geneexpression (Weber et al., 2007).

Axl is a 140-kDa protein, activated either withgrowth-arrest-specific gene 6 (Gas6) or homophilicinteractions and activates different signaling moleculeslike phosphatidylinositol 3-kinase, Akt, Src, extracellu-lar signal-regulated kinase and nuclear factor kappaB(Goruppi et al., 1997; Lee et al., 2002; Vajkoczy et al.,2006). Overexpression of Axl can transform fibroblastseven in the absence of a ligand (Burchert et al., 1998).Axl is known to induce cell survival (Melaragno et al.,2004; van Ginkel et al., 2004), proliferation (Stenhoffet al., 2004; Sainaghi et al., 2005), stimulation of cellmigration (Fridell et al., 1998) and cell–cell adhesion(McCloskey et al., 1997). Moreover, an increasedexpression of Axl is associated with invasion, metastasis,angiogenesis, and is found in metastatic colon, prostatecarcinoma, gastric and endometrial cancers, breastcancers, lung cancers and sarcomas (Hafizi and Dahl-back, 2006). Axl is transcriptionally regulated by Sp1/Sp3 transcription factors and further controlled by CpGisland methylation (Mudduluru and Allgayer, 2008).Additionally, overexpression of MZF1 transactivatesAxl gene expression and induces migration, invasionand in vivo metastasis formation (Mudduluru et al.,2010). However, little is known about the epigeneticregulation of Axl and especially its post-transcriptionalregulation by microRNAs (miRNAs).

MiRNAs are a class of B22 nt endogenous RNAs,which can be expressed in a cell and tissue-specificmanner, influencing mRNA stability and translation.They control a wide range of biological functions suchas cellular proliferation, differentiation and apoptosis(Zhao and Srivastava, 2007; Erson and Petty, 2008;

Received 29 August 2010; revised 18 December 2010; accepted 4 January2011; published online 14 February 2011

Correspondence: Professor H Allgayer, Department of ExperimentalSurgery/Molecular Oncology of Solid Tumors (Collaboration UnitGerman Cancer Research Center-DKFZ-Heidelberg), MedicalFaculty Mannheim, Ruprecht-Karls-University of Heidelberg,Mannheim 68167, Germany.E-mail: [email protected]

Oncogene (2011) 30, 2888–2899& 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11

www.nature.com/onc

Eulalio et al., 2008). Recent reports showed strongevidence that miRNAs can act as oncogenes or tumorsuppressors, having key roles in cancer initiation andprogression (Cho, 2007; Tili et al., 2007). In the attemptto understand which mechanisms underlie abnormalmiRNA expression in cancer, an increasing number ofstudies have investigated how miRNAs are regulated,and it is now widely accepted that miRNAs undergo thesame regulatory mechanisms as any other classicalprotein-coding genes, including epigenetic regulation(Valeri et al., 2009). Recent papers have demonstratedthat the epigenetic regulation of miRNAs in cancer is awidespread phenomenon, for example, as shown formiR-9-1, miR-107, miR-127, miR-193a, miR-137, miR-342, miR-203, miR-34b/c and miR-1 (Lujambio andEsteller, 2009; Valeri et al., 2009).

All this emerging evidence prompted us to design astudy on several non-small cell lung cancer (NSCLC),breast cancer (BRC) and colorectal cancer (CRC) celllines, as well as in a case-series of resected tissues ofNSCLC patients, to test the hypothesis and the possibleimpact of Axl regulation by miRNAs.

Results

An evolutionary conserved target sequence for miR-34aand miR-199a/b in the 30-UTR of AxlThe 1868 nt 30-UTR of Axl was screened for comple-mentary seed sequences of known miRNAs via abioinformatic search. A 100% match target sequencefor miR-34a at nts 24–50, and for miR-199a/b at nts25–56, was found (Figure 1a). The minimum free energypredicted for hybridization with the Axl 30-UTR andmiR-34a or miR-199a/b at their site is DG –31.8,DG –20.7 and DG –21.7 kcal/mol, respectively, deter-mined by mFold analysis (Supplementary Figure S1),this being consistent with authentic miRNA targeting(Doench and Sharp, 2004). Comparing the humansequence for interspecies homology, we found that themiR-34a and miR-199a/b target sequences at nts 24–50and at nts 26–54 of the Axl 30-UTR are highly conservedamong species (Figure 1b).

The Axl 30-UTR is a target for miR-34a and miR-199a/bGiven the results from homology search across species,we asked whether the 30-UTR of Axl is a functionaltarget of miR-34a and miR-199a/b. We cloned areporter plasmid driven by the cytomegalovirus basalpromoter, harboring the 1834 nt 30-UTR of Axl at the30-position of the luciferase-reporter gene (Axl 30-UTR).We also cloned a shorter fragment (1741 nt 30-UTR) byremoving 93 nt at the 50-end of the Axl 30-UTR (Axl del-30-UTR), which does not contain the seed sequences formiR-34a and miR-199a/b. As miR-199 expression wascomparatively low in the screened cells (SupplementaryFigure S4b), both of the constructs were transfected intohigh miR-34a-expressing cells (H460, MCF-7 and Colo320). Axl del-30-UTR significantly induced luciferaseactivity when compared with the complete Axl 30-UTR

construct (Figure 2a, left column panel) (*Po0.05).Furthermore, co-transfection of the Axl 30-UTR alongwith AM-miR-34a significantly induced luciferase activ-ity when compared with the corresponding control-miR(Figure 2a, middle column panel) (*Po0.05). Addition-ally, PM-miR-34a or PM-miR-199a/b were co-trans-fected along with the Axl 30-UTR-luciferase constructinto H1299, MDA-MB-231 and Rko cells, which lackendogenous expression of these three miRs, significantlyreducing the luciferase activity when compared withcontrol-miR (Figure 2a, right column panel) (*Po0.05).Taken together, these data suggest that the 30-UTR ofAxl is a functional target for miR-34a and miR-199a/b.

MiR-34a and miR-199a regulate Axl gene expressionand competes for binding sequenceTo corroborate the reporter assay results on miR-34-and miR-199-inhibited Axl regulation at the mRNA andprotein level, H1299, MDA-MB-231, HCT116 and Rkocells were transfected with control-miR, miR-34a ormiR-199a. As miR-199a and miR-199b have similarseed sequences and binds with more or less equal freeenergy to the Axl 30-UTR, miR-199a was used forfurther transfection experiments. qRT–PCR resultsshowed that both the miRs (miR-34a and miR-199a)significantly reduced Axl-transcript levels (Figure 2b).Furthermore, western blot analysis confirmed the down-regulation of Axl protein amounts (Figure 2c). Simi-larly, other known targets of these miRs (miR-34a:c-Met and Notch 1; miR-199a: HIF1a) were screenedthrough western blot analyses and observed the down-regulation of the respective molecules by the respectivemiRs (Supplementary Figure S2) (Li et al., 2009; Yeligaret al., 2009). However, transfection efficiency of miR-34a/miR-199a and expression were measured with qRT–PCR (Supplementary Figure S3a). Taken together,our results suggest that both the miRs are regulatingAxl at the transcript level, which downregulates proteinamounts.

Additionally, experiments were carried by transfect-ing AM-miR-34a, which binds to endogenous miR-34aand thereby antagonizes its activity. As expected, inA549 cells Axl protein amounts were increased, but noAxl protein band was detected in MCF-7 and Colo 320cell lines (Figure 2d). Transfection efficiency of the AM-miR-34a and expression levels were measured withqRT–PCR (Supplementary Figure S3b). This result isnot surprising with MCF-7 and Colo 320 cells, as weknow that Axl is also epigenetically regulated by hyper-methylation of its promoter (Mudduluru and Allgayer,2008).

As miR-34a and miR-199a seed sequences arepartially overlapping, additional experiments wereperformed transfecting Rko and H1299 cells with PM-miRs, or with the combination of PM-miRs and AM-miRs, and by the analysis of Axl protein expression bywestern blot. The results showed that silencing of eithermiR-199a or miR-34a through the respective AM-miRmutually enhances the miR-34a/miR-199a efficiency indownregulating Axl, respectively (Figure 3). Comparedwith miR-34a, miR-199a showed to downregulate Axl

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2889

Oncogene

expression less efficiently (Figures 2c and 3). Moreover,an evident downregulation of Axl protein was observedafter transfecting cells with AM-miR-199a and PM-miR-34a, while AM-miR-34a and PM-miR-199a co-transfection had poorly significant effects. These resultsindicate that miR-34a and miR-199a are competing fortheir seed sequence within the Axl 30-UTR, miR-34abeing more efficient than miR-199a in controlling Axlexpression.

Axl-mRNA and Axl protein correlates inversely withmiR-34a expression and invasion of NSCLC, BRCand CRC cell linesThe expression of Axl-mRNA, Axl protein, miR-34aand miR-199a/b were screened in a panel of NSCLC,BRC and CRC cell lines and correlated with cellinvasive ability. MiR-34a, miR-199a/b and Axl endo-genous expression was measured by qRT–PCR (Sup-plementary Figures S4a, b and S5a). MiR-199a/bendogenous expression was not detected in most of thescreened cancer cells (data not shown). Axl proteinamounts were estimated through western blot analysisand densitometry as the ratio Axl/b-actin (Supplemen-tary Figure S5b, c). The invasive ability of NSCLC,BRC and CRC cells was measured by Matrigel assayand plotted as the percentage of invading cells after 15 h(Figure 4a). A statistically significant positive correla-tion between Axl-mRNA, or Axl protein expression and

the invasion of NSCLC, BRC and CRC cells, respec-tively, was found (RS¼ 0.64, P¼ 0.001; RS¼ 0.74,Po0.001). Moreover, an inverse correlation betweenmiR34a expression and the invasiveness of NSCLC,BRC and CRC cells was found (RS¼�0.45, P¼ 0.03).Axl protein and miR-34a expression correlated inverselywith a statistically positive association, when comparedwith all cell lines (RS¼�0.38, P¼ 0.07) (Figures 4b–e).However, specifically NSCLC cells showed statisticallysignificant correlations (Supplementary Figure S6a–c).Correlations were evaluated by the Spearman’s rankcorrelation method. Taken together, from these resultswe can hypothesis that Axl is positively regulatinginvasion, and miR-34a is inhibiting Axl expression andinvasion.

MiR-34a and miR-199a inhibits in vitro migrationand invasion and in vivo distant metastasisTo further investigate the ability of miR-34 andmiR-199 to regulate diverse phenomena of the meta-static cascade, a wound-healing assay was performedwith H1299 cells by transfecting control-miR, pre-miR-34a or pre-miR-199a (Figure 5a). Furthermore, amigration assay was performed with trans-well chambersby transfecting control-miR, pre-miR-34a or pre-miR-199a, into H1299 and Rko cells (Supplementary FigureS7). Moreover, Matrigel invasion assays were performedwith highly invasive H1299, MDA-MB-231 and Rko

Homo sapiens:Pan troglodytes:Macacamulatta:Mus musculus:Rattus norvegicus:Cavia porcellus:Oryctolagus cuniculus:Erinaceus europaeus:Canis familiaris:Equus caballus:Bos taurus:

miR-199amiR-199b

miR-34a

Axl3 UTR 1868 nt---TGA

(nt 24-50)miR-34a

(nt 25-56)miR-199a

’

Figure 1 miR-34a and miR-199a/b target sites reside in close proximity with the Axl 30-UTR, and are highly conserved across 11species. (a) The location of the putative miR-34 and miR-199a/b target sites are shown with underlined and bold letters, respectively,unpaired bases are shown either above or below the duplexes. (b) Comparison of nucleotides between the miR-34 and miR-199a/b seedsequence and the target Axl site within the 30-UTR across 11 species.

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2890

Oncogene

cells (Figure 4a) transfected with control-miR, pre-miR-34a or pre-miR-199a (Figure 5b). In all experiments,miR-34a and miR-199a significantly reduced migrationand invasion when compared with control-miR trans-fected cells (*Po0.05).

To determine the ability of miR-34a and miR-199a insuppressing the ability of cancer cells to metastasizein vivo, a chorion-allantoic-membrane (CAM) assay wasperformed with H1299 and Rko cells transfected withcontrol-miR, pre-miR-34a, or pre-miR-199a. The num-ber of metastatic H1299 and Rko cells into embryonicchicken liver and lungs was dramatically reduced whencompared with control-miR transfected cells (Figures 6cand d) (*Po0.05). Additionally miR-34a and miR-199ainhibited the primary tumor growth in the upper CAM,in a separate experiment (Figures 6a and b) (*Po0.05).To check Axl and miR-expression in parallel to the

migration and invasion process, RT–PCR and westernblot quantifications were additionally performed 8 daysafter transfection. The results showed a reduction inmiR-expression after day 8 (o30-fold) as comparedwith the second day of transfection, but no significantchanges in Axl protein amounts (data not shown).Taken together, these results suggest that miR-34aand miR-199a are inhibiting several different steps ofmetastasis, migration, invasion and the formation ofin vivo distant metastasis.

5-Aza treatment reactivates miR-34a, miR-199a and bexpression and inhibits Axl proteinTo further differentiate the epigenetic mechanism ofregulation of miR-34, miR-199 and Axl, highly invasive(H1299, MDA-MB-231 and Rko), and less invasive cells(H520, Colo 206f and Geo) were treated with a selective

Per

cen

t lu

cife

rase

acti

vity

Per

cen

t lu

cife

rase

acti

vity

Per

cen

t lu

cife

rase

acti

vity

1: Axl 3'UTR2: Axl del-3'UTR

050

100150200250

1 2

0

200

400

600

800

1 2

0

100

200

300

1 2

H460

MCF-7

Colo 320

H1299

020406080

100120

Con 34a 199a 199b

020406080

100120

020406080

100120

PM-miR

MDA-MB-231

Con 34a 199a 199bPM-miR

Con 34a 199a 199bPM-miR

Rko

Con: Control-miR (Scrambled)

04080

120160200

050

100150200250

04080

120160

1 2

1 2

1 2

H460

MCF-7

Colo 320

1: Control-miR2: AM-miR-34a

04080

120

Axl

pro

tein

(%)

1 2 3

04080

120

1 2 3

Axl

pro

tein

(%)

Axl

pro

tein

(%)

0

40

80

120

1 2 3

04080

120

Axl

pro

tein

(%)

1 2 3

H1299

Axl

�-Actin

Axl

�-Actin

MDA-MB-231

RkoAxl

�-Actin

HCT116Axl

�-Actin

1: Control-miR 2: PM-miR34a 3: PM-miR199a

0.0

0.4

0.8

1.2

H1299

Rel

ativ

e A

xl-m

RN

Aex

pre

ssio

n 1.2

0.0

0.4

0.8

Rko

1 2 3 1 2 30

0.4

0.8

1.2

1 2 3

MAD-MB-231

1: Control-miR 3: PM-miR199a

Axl

�-Actin

A549 MCF-7 Colo 320

1: Control-miR; 2: AM-miR-34a

130 kDa

43 kDa

130 kDa

43 kDa

130 kDa

43 kDa

130 kDa

43 kDa

130 kDa

43 kDa

212121

321

321

321

321

2: PM-miR34a

Figure 2 miR-34a and miR-199a/b target the Axl 30-UTR and regulate Axl expression. (a) Luciferase-reporter assays of Axl 30-UTR(complete 30-UTR) and Axl del-30-UTR (seed sequence of miR-34a and -199a/b, 50 region of Axl-30-UTR removed) in H460, MCF-7and Colo 320 (left column) or Axl 30-UTR with co-transfection of either with control-miR or AM-miR-34a in H460, MCF-7 and Colo320 (middle column) or Axl 30-UTR with co-transfection of either with control-miR or PM-miR-34a, PM-miR-199a/b in H1299,MDA-MB-231 and Rko (right column) as indicated. Percent luciferase activity was calculated either with Axl 30-UTR or control-miRsamples set as 100%. Each bar represents values of quadruplicates (Pp0.05) (s.d. is small). H1299, MDA-MB-231, HCT116 and Rkocells were transfected either with miR-34a or miR-199a. After 48 h, total RNA and protein was isolated, and (b) Axl expression levelswere evaluated by RT–PCR (Po0.05), (c) Axl protein amounts were estimated by western blot analysis and quantified bydensitometry, density ratio of Axl/b-actin is represented as a bar graph. In control-miR transfected samples, Axl protein amounts wereset as 100% and relative Axl protein amounts were calculated in PM-miR transfected samples and represented as indicated. Westernblots were represented adjacent to the bar graphs, respectively. (d) A549, MCF-7 and Colo 320 were transfected either with control-miR or AM-miR-34a and after 48 h protein was isolated and western blot analysis for Axl and b-actin was performed.

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2891

Oncogene

inhibitor of DNA methyltransferases (5-aza). Cells weretreated for 5 days with 2 mM 5-aza by changing the freshmedium with drug and DMSO control. Interestingresults were observed regarding Axl gene expressionafter 5-aza treatment. Axl-mRNA levels were signifi-cantly downregulated in highly invasive cells, and theopposite effect was observed in less invasive cells(Figure 7b, top row panel). The fact that 5-azatreatment induced Axl gene expression in less invasivecells is consistent with our previous results (Mudduluruand Allgayer, 2008). However, Axl protein amounts

were drastically reduced after 5-aza treatment(Figure 7a). This suggests that Axl regulation iscontrolled at transcriptional level by CpG hyper-methylation and also post-transcriptionally by miRs.5-Aza treatment induced the constitutive expression ofmiR-34a (Figure 7b, second row panel), miR-199a(Figure 7b, third row panel) and miR-199b (Figure 7b,last row panel). In some cell lines (MDA-MB-231,H520, Geo and Colo 206f), mixed results were observed,regarding miR-expression: either no significant changein expression of any one/two miRs, or a significantinduction of one particular miR was observed(Figure 7b). This suggests that, apart from CpGmethylation, other mechanisms could control the expres-sion of these miRs, possibly in a cell-line specific manner.

Hyper-methylation of miR-34a, miR-199a/b and AxlpromotersTo understand the mechanisms behind the downregula-tion of miRs (miR-34a (MI0000268), miR-199a1(MI0000242), miR-199a2 (MI0000281) and miR-199b(MI0000282) in solid cancer cells (NSCLC, BRC, andCRC), DNA was isolated and subjected to bisulfiteconversion and to a subsequent methylation-specificPCR with ‘methylated’ and ‘unmethylated’ pairs ofprimers designed in the CpG island promoter region ofmiR-34a, miR-199a and b, and also Axl (SupplementaryFigure S8). As expected, miR-199a and b promoterswere hyper-methylated in most of the screened cells,correlating with their expression. miR-34a expressionlevels were high in A549 and H460; particularly in thesetwo cell lines, miR-34a promoter hypo-methylation wasobserved. Axl promoter hyper-methylation was des-cribed by us previously (Mudduluru and Allgayer, 2008)

H1299Rko

1: Contol-miR2: PM-miR34a3: PM-miR199a

0

40

80

120

1Axl

Pro

tein

(%

)

0

40

80

120

Axl

Pro

tein

(%

)

65432 1 65432

4: PM-miR34a+PM-miR199a5: PM-miR34a+AM-miR199a6: AM-miR34a+PM-miR199a

654321 654321

Figure 3 MiR-34a and miR-199a compete for seed sequences. Axlprotein amounts were estimated by western blot analysis andquantified by densitometry, density ratio of Axl/b-actin isrepresented as a bar graph. In control-miR-transfected samples,Axl protein amounts were set as 100%, and relative Axl proteinamounts were calculated in PM-miR or in combination (PM-miRand AM-miR) transfected samples.

0.0

0.8

1.6

Rko

HC

T11

6

Co

lo32

0H

T-2

9

WiD

r

Co

lo20

6f

SW

480

HC

T15

Geo

Inva

sio

n (

%)

01234

H12

99C

alu

-1H

522

LX

F28

9H

596

Cal

u-3

H52

0S

K-M

ES

1A

549

H13

95H

460 0

61218

MC

F-7

MD

A-

MB

-231

NSCLCa

b c d e

BRC CRC

Axl-mRNA vs invasion miR-34a vs invasion miR-34a vs Axl-proteinAxl-protein vs invasion

Log[miR-34a expression]

Inva

sio

n a

bili

ty R=-0.45; p=0.03

Axl-mRNA expression

Inva

sio

n a

bili

ty

R=0.64; p=0.001

1412

10

86

42

0

1412

10

86

42

0

14

12

108

64

2

00 2 4 6 0

0.0

0.5

1.0

1.5

2.0

2.5

3.0

2 4 60 1 2 3 0 1 2 3Axl-protein expression

Inva

sio

n a

bili

ty

R=0.74; p<0.001

Log[miR-34a expression]

Axl

-pro

tein

exp

ress

ion

R=-0.38; p=0.07

Figure 4 Invasive ability of NSCLC, BRC and CRC cell lines and correlations between invasion, miR34a and Axl endogenousexpression. (a) Invasive ability of the cells evaluated by Matrigel assays, data are expressed as percentage of invaded cells comparedwith non-invading cells. Correlations between (b) Axl-mRNA expression vs invasion, (c) Axl protein vs invasion, (d) miR34a-expression vs invasion, (e) miR-34a expression vs Axl protein in a of panel of NSCLC, BRC and CRC cell lines (P-value is from aSpearman’s correlation test).

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2892

Oncogene

correlated with Axl expression in some of the screenedcells (H520, Colo 206f, Geo), but not in all of them,suggesting the biological relevance of the Axl-regulatingmechanisms besides epigenetics (Figure 7b, top panel;Figure 8).

MiR-34a, miR-199a and miR-199b are frequentlydownregulated and co-expressed in NSCLC patientsTo investigate an in vivo relevance of the mechanismsfound for cancer patients and their tumor diseases, miR-34a, miR-199a and miR-199b expression was tested inNSCLC samples and in the corresponding normal lungtissues. Results were analyzed adopting a cutoff value of0.75 (that is, 1.5-fold) for expression. Concerning miR-34a, the results showed a significant downregulation in24 out of 44 patients (54.5%; DDCt o�0.75), while 12(27.3%) patients had a ratio 40.75, that is, miR-34asignificantly upregulated. Concerning miR-199a, theresults showed a significant downregulation in 19 outof 44 patients (43.2%), while 13 (29.5%) patients, miR-199a was upregulated. Finally, miR-199b was found tobe downregulated in 23 out of 44 patients (52.3%), whilein 9 cases (20.5%) this miR was upregulated. Rest of thepatients for these three miRs did not show significantchanges in their expression when compared with theirrespective normal tissues.

The three miRs were found to be significantly co-regulated (see Figures 9a–c) and a lower expression ofall three was significantly associated with a squamousNSCLC histotype (Figure 9d). No other significantcorrelations between miR levels and patients’ character-istics such as gender, age, tumor size, lymph nodal statusand clinical stage were found (not shown). Survival

analysis, dividing the patients between these with up-regulated and those with down-regulated or equal,showed a positive association towards a longer survivalfor the patients with an upregulation of miR-34a(P¼ 0.07), while miR-199a and 199b were not correlatedwith patients’ prognosis (Figure 9e). Furthermore, animmunohistochemical quantification of Axl protein wasperformed in those tumor specimens with miR-34anormal/tumor ratio o�2 or 42 (that is, stronglydownregulated and upregulated). As a result, we foundno differences of Axl expression in terms of bothintensity and percentage of positive cells, but interest-ingly a more frequent cytoplasmic staining was observedin those tumors with strong miR-34 downregulationas compared with upregulated tumors (62 vs 14%, seeFigure 9f).

Discussion

The major finding of this study is that three micro-RNAs, miR-34a, miR-199a, and miR-199b can inhibitthe expression and functions of Axl tyrosine kinase,thereby impairing migration, invasion and formation ofdistant metastasis of cancer cells (Vajkoczy et al., 2006;Mudduluru et al., 2010). As 5-aza treatment was shownto induce the constitutive expression of miR-34a, miR-199a, and miR-199b and to inhibit Axl proteinexpression, we found that miR-34a and miR-199a/bwere frequently methylated and that their expressionlevels significantly inversely correlated with invasivecapacity and Axl expression. Our study extends theresults of others showing the miR-34a and miR-199a2regulation is controlled by methylation (Lodygin et al.,

Control-miR (0 h)

2.2

2.12

2.12

miR-34a (8 h)

1.35

0.95

1.14

miR-199a (8 h)

1.14

0.93

1.06

0.81

0.7

0.84

Control-miR (8 h)

020406080

100120

Co

ntr

ol-

miR

miR

-34a

miR

-199

a

Cel

l mig

rati

on

(%

) 8 hours

* *

p=.04 p=.02

H1299

048

1216

Co

ntr

ol-

miR

miR

-34a

miR

-199

a

Inva

sio

n(%

)

Rko

00.20.40.6

Inva

sio

n(%

)

MDA-MB-231

048

12

Co

ntr

ol-

miR

miR

-34a

miR

-199

a

Inva

sio

n(%

)

Co

ntr

ol-m

iR

miR

-34a

miR

-199

a

*** * *

Figure 5 MiR-34a and miR-199a inhibits the migration and invasion in vitro. (a) H1299 cells were transfected with PM-miR-34a andPM-miR-199a. After 48 h wounds were created with a yellow tip, pictures were taken and again after 8 h. Wounds were measured inthree different places and mean wound distance of control-miR transfected set as 100%. PM-miR-34a and PM-miR-199a transfectedsamples were calculated and represented as a graph (*Po0.05). (b) H1299, MDA-MB-231 and Rko cells were transfected with control-miR or PM-miR-34a or PM-miR-199a. After 48 h cells were plated on top of the Matrigel-coated Boyden chambers, invaded cells weremeasured as described in Materials and methods. Data are represented as the percentage of invading cells, as mean±s.d. of fourreplicates (*Po0.05).

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2893

Oncogene

2008; Toyota et al., 2008; Cheung et al., 2010). More-over, to the best of our knowledge this is the first reportto show the methylation of miR-199a1 and miR-199bpromoters.

In our work, by in silico analysis, we have identifiedmiR-34a and miR-199a/b as potential candidates cap-able of targeting the Axl oncogene. By co-transfectingmiRNAs and Axl 30-UTR either with or without seedsequences for the predicted miRNAs, we have shownthat Axl is a true target for the miR-34a and miR-199a/b. Moreover, exogenous expression of each miR-34a ormiR-199a resulted in a strong decrease in Axl protein,whereas blockade of endogenous miRNAs (by anti-miRs) led to a twofold increase of Axl expression inA549. The lack of a specific Axl band observed forMCF-7 and Colo 320 cells was not surprising, as Axltranscriptional regulation is epigenetically controlledby CpG methylation in these particular cell lines(Mudduluru and Allgayer, 2008).

5-Aza treatment increased Axl-mRNA expression inH520, Geo and Colo 206f (Mudduluru and Allgayer,2008), but not in H1299, MDA-MB-231 and Rko cells.H1299, MDA-MB-231 and Rko cells are more invasiveand have high endogenous Axl expression. In particular,these cells might be bypassed the epigenetical regulation,where as H520, Geo and Colo 206f cell lines are lessinvasion and no Axl expression. Axl-mRNA levels were

induced in H520, Geo and Colo 206f cells, since it iscontrolled by CpG methylation. However, proteinamounts were significantly reduced after 5-aza treatmentin all of these cells (Vajkoczy et al., 2006; Mudduluruand Allgayer, 2008; Mudduluru et al., 2010). In general,it implies different stages of controlling mechanisms of agene, especially of an oncogene. Moreover, miR-34a andmiR-199a2 are reported that they are epigeneticallycontrolled by CpG methylation (Lodygin et al., 2008;Cheung et al., 2010). 5-Aza treatment significantlyinduced the expression of either one or all of the miRs(miR-34a, miR-199a and b) in the screened cell lines.Epigenetic regulation by CpG methylation of thesemiRs could be cell line specific or other epigeneticmechanisms and specific transcriptional factors mightinterfere with the expression. Bisulfite conversion andmethylation-specific PCR supported the existing dataand also revealed that miR-34a, miR-199a1, miR-199a2and miR-199b promoters were hyper-methylated andinversely correlated with their expression in a panel ofcancer cells (Lodygin et al., 2008; Garzia et al., 2009;Cheung et al., 2010). Our data shows in particular thatmiR-199a1 and miR-199b promoters were hyper-methylated in NSCLC, BRC and CRC cell lines. CpGmethylation might account for the loss of these threemiRs in a substantial portion of carcinomas (Bommeret al., 2007; Chang et al., 2007; Tazawa et al., 2007;

0

200

400

600

800

Control-miR miR-34a miR-199a

Met

asta

size

d c

ells

into

live

r

* *

0.00.20.40.60.81.01.2

a

c

b

d

Control-miR

Tu

mo

r w

eig

ht

(mg

)

***

0

20

40

60

80

100

120

Met

asta

size

d c

ells

into

lun

g

* *

0.00

0.01

0.02

0.03

0.04

Control-miR

Tu

mo

r w

eig

ht

(mg

)

0

500

1000

1500

2000

2500M

etas

tasi

zed

cel

lsin

to li

ver

0

1000

2000

3000

4000

5000

Met

asta

size

d c

ells

into

lun

g

H1299 Rko

H1299 Rko

H1299 Rko

** **

** **

miR-199amiR-34a

Control-miR miR-199amiR-34a

Control-miR miR-199amiR-34a

Control-miR miR-199amiR-34a

miR-199amiR-34a

Figure 6 MiR-34a and miR-199a reduces tumor growth and inhibits metastasis in vivo. Chicken embryo metastasis (CAM) assay wasperformed with H1299 and Rko cells transfected with control-miR or PM-miR-34a or PM-miR-199a. Ten eggs were included for eachgroup. (a, b) Upper CAM tumors were removed and weights were measured and represented as indicated. (c, d) Genomic DNA wasanalyzed by real-time Alu-PCR to determine the number of metastasized cells into liver and lung (right), ±s.d. of eight replicates(*Po0.05 and **Po0.01).

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2894

Oncogene

Lodygin et al., 2008; Cheung et al., 2010). Our datashow that Axl-targeting miRs are epigenetically silencedin cancer cell lines/tumors, and that miR-34a and miR-199a/b are regulating Axl expression at mRNA levels bydegrading it and also inhibiting translation, which iscorroborating the Axl overexpression in different typesof cancers (Hafizi and Dahlback, 2006). As Axl andmiR-34a have opposite functions; their endogenousexpressions were quantified and correlated withNSCLC, BRC and CRC invading capacity. Axl-mRNAand Axl protein expression significantly inverselycorrelated with the miR-34a expression, and cellinvading capacity positively correlated with Axl proteinamounts and inversely with miR-34a expression. Ingeneral, our findings strongly support the existingliterature that Axl is an oncogene and miR-34a is atumor suppressor (Hafizi and Dahlback, 2006; Vajkoczyet al., 2006; Bommer et al., 2007; Welch et al., 2007;Mudduluru et al., 2010).

Previous studies reported that miR-34a can inhibitcell cycle (CCNE2, CDK4, CDK6, Cyclin E2 andE2F5), anti-apoptotic protein (BCL2) and invasion

(MET) inducing genes (Bommer et al., 2007; Changet al., 2007; He et al., 2007; Raver-Shapira et al., 2007).Additionally, 199a and 199b can inhibit cell prolifera-tion by targeting cell proliferation inducers like IKKb,HES1, Cyclin D1 and C-Myc (Chen et al., 2008; Garziaet al., 2009). miR-34a is a p53 target gene thatpresumably mediates induction of apoptosis, cell cyclearrest and senescence by p53. Subsequently, Axl isknown to induce migration, invasion and distantmetastasis (Vajkoczy et al., 2006; Mudduluru et al.,2010). We now demonstrate that miR-34a and miR-199a are potent inhibitors of migration, invasion(H1299, MDA-MB-231 and Rko) and tumor growthand, in vivo distant metastasis by CAM assay (H1299and Rko). Thus, the inhibition of migration andinvasion by these miRs might in part be mediated vianegative regulation of Axl. However, the present workon miR-34a and miR-199a does not exclude otherimportant mechanisms in terms of known and/orunknown additional targets of these miRs, especiallyin the context of tumor progression, invasion andmetastasis. Towards this end, in the present work we

02468

1012

ba

Rel

ativ

e A

xl-m

RN

Aex

pre

ssio

n

H1299 MDA-MB-231

Rko H520 Geo Colo 206f

0123456

0.1

1

10

100

1000

0

4

8

12

16

Rel

ativ

e m

iR-3

4aex

pre

ssio

n

H1299 Rko H520 Geo Colo 206f

Rel

ativ

e m

iR-1

99a

exp

ress

ion

(L

og

sca

le)

H1299 Rko H520 Geo Colo 206f

Rel

ativ

e m

iR-1

99b

exp

ress

ion

H1299 Rko H520 Geo Colo 206f

1: DMSO; 2: 5 Aza (2 �M)

MDA-MB-231

MDA-MB-231

MDA-MB-231

Axl

�-Actin

�-Actin

�-Actin

Axl

Axl

Colo 206f

H1299

MDA-MB-231

RkoGeo

H520

1: DMSO; 2: 5 Aza (2 �M)

130 kDa

43 kDa

130 kDa

43 kDa

130 kDa

43 kDa

21 21

21 21

21 21

212121212121

212121212121

212121212121

212121212121

Figure 7 5-Aza-dC (5-aza)-treatment for 5 days induced the expression miR-34a, miR-199a, miR-199b and deregulated Axl geneexpression. H520, H1299, MDA-MB-231, Colo 206f, Geo and Rko cells were treated with 5-aza (2mM) for 5 days and total RNA andprotein was isolated. (a) Western blot analysis was performed for Axl and b-actin as indicated, (b) Axl-, miR-34a-, 199a/b-expressionlevels were evaluated by RT–PCR (s.d. is small, thus hardly visible).

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2895

Oncogene

have confirmed that c-Met and Notch-1, and alsoHIF1a, are targets of miR-34a and miR-199a, respec-tively, molecules that are known to be highly relevantfor tumor formation and progression, and also metas-tasis. The targeting of these mRNAs by miR-34a/199ahas been shown by Li et al. (2009) and Yeligar et al.(2009), in glioblastoma and liver sinusoidal endothelialcells, respectively. These and our results, together withthe very likely notion that further miR-34a/199a targetswill be discovered soon, we consider it very likely thatthese two miRs are capable of regulating a wholeprogram of concerted regulation of progression- andmetastasis-related mRNAs.

To confirm some of the observations obtained in cellline experiments, an expression analysis was performedin primary tumor specimens from a case-series ofconsecutively resected NSCLC patients. As a result ofqPCR experiments, miR-34a, miR-199a and miR-199bwere found frequently downregulated in the tumorswhen compared with the corresponding normal tissues,and a significant co-expression in the tumor tissuesamong the three miRs was observed. While nosignificant correlation between miR expression levelsand patients’ clinico-pathological characteristics (gen-der, age, tumor size, lymph nodal status and clinicalstage) was found, survival analysis indicated a positiveassociation towards a longer survival for those patientswith miR-34 upregulation. This is in line with recentresults indicating miR-34 as a prognostic factor inNSCLC (Gallardo et al., 2009), and further confirm itsputative role as a tumor suppressor miR in lung cancer.Interestingly, a comparably lower level of regulation(that is, expression) in squamous, as compared with

non-squamous tumors, has been observed for all thethree miRs. Additionally, the quantification of Axlprotein by immunohistochemistry confirmed a differ-ential pattern of expression depending on the status ofmiR-34a regulation, being Axl more frequently localizedin the membrane (that is, active) in NSCLC patientswith miR-34a downregulation. Hypothetically, thepresence of Axl in the cytoplasm (more frequent intumors with high miR-34a) could be due to proteosomaldegradation or other proteolytic degradation (forexample by ADAM10 or maybe other functionalchanges caused by specific miRs in tumors). We didnot extend specific data on this aspect at thispoint of time in our present study (O’Bryan et al.,1995; Budagian et al., 2005), but consider it as animportant aspect that those patients with high miR-34aexpression appear to have less functional (membrane-bound) Axl protein as suggested by IHC, the molecularmechanism leading to this still being in need to beinvestigated.

In conclusion, this evidence highlights a pivotal rolefor miR-34a and miR-199a/b in various aspects oftumorigenesis like tumor growth, migration, invasion,and in vivo distant metastasis in cancer cell linesespecially through Axl regulation.

Materials and methods

Cell lines, cultures and drugsEleven human NSCLC cell lines (Calu-1, H520, SK-MES-1,H596, Calu-3, H522, H1395, H1299 H460, LXF289, A549 andH520), two BRC cell lines (MCF-7, MDA-MB-231) and

bp: Base pairs; Mr: Marker; M: Methylated; U: Unmethylated1: H1299; 2: Calu2; 3: H522; 4: Calu3; 5: H520; 6: SKMES1; 7: A549; 8: H1395; 9: H460; 10: MDA-MB-231; 11: MCF-7; 12: Rko; 13:HCT116; 14: SW480; 15: HT-29; 16: WiDr; 17: Geo; 18: Colo206f

miR-199a II

miR-199a I

miR-199b

Mr

miR-34a1

Axl

bp400

100

400

100

bp

300100

bp

300100

bp

400

100

bp

18171615141312111098765432

1 18171615141312111098765432

1 18171615141312111098765432

1 18171615141312111098765432

UMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUM

UMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUM

1 18171615141312111098765432

UMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUM

UMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUM

UMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUMUM

Mr

Mr

Mr

Mr

Figure 8 Methylation status of Axl, miR-34a and 199a/b promoters in cancer cells. Gel electrophoresis of PCR products obtainedfrom methylation-specific PCR for miR-34a, miR-199a, miR-199b and Axl in a panel of cancer cell lines.

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2896

Oncogene

eight CRC cell lines (Rko, HCT116, SW480, Colo 320, HT-29,HCT15, WiDr and Colo 206f) were purchased from AmericanType Culture Collection (ATCC, Manassas, VA, USA), andgrown at 37 1C with ATCC recommended media supplementedwith 10% fetal calf serum. Geo (CRC cell line) was a gift fromD Boyd (MD Anderson Cancer Center, Houston, TX, USA),and grown as similar to the other cell lines with DMEM.Original stock solutions of 5-aza-20-deoxycytidine (5-aza-dC,Sigma Chemical Co., St Louis, MD, USA) at a concentrationof 4mM was stored at �20 1C and freshly dissolved in culturemedium before use.

Patients and samplesFresh snap-frozen surgical specimens of tumor tissues and ofthe corresponding normal specimens from 44 NSCLC patientscompletely resected between 2005 and 2006 at the San LuigiHospital (Orbassano, Italy) were consecutively collected. Themain patients’ characteristics are reported in SupplementaryTable S1. None of the patients received pre-operatory chemo/radiation therapy. All cases were reviewed and classifiedaccording to the WHO classification by one of the investiga-tors (MP), using anonymous samples; none of the researchersconducting gene expression and statistical analyses had accessto disclosed clinical-pathological data. The study wasapproved by the Institutional Review Board of the UniversityHospital.

Construction of 30-UTR-luciferase plasmids and reporter assaysThe full-length 30-UTR of Axl (1834 nt) was amplified usingcDNA from H1299 and cloned into the HindIII-site ofpMIR (Ambion, Austin, TX, USA), checked for orientation,sequenced and named Axl 30-UTR. Axl del-30-UTR (withdeleted seed sequences of miR-34a and miR-199a/b) wasamplified using the Axl 30-UTR luciferase construct as atemplate. Cloning primers are provided in SupplementaryTable S2. For reporter assays, cells were co-transfected usinglipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) eitherwith 1 mg of luciferase construct and pRL-TK (50 ng, RenillaLuciferase; Promega, Madison, WI, USA) or along with 50 nMof control-miR or PM/AM miR. pRL-TK was co-transfectedand luminescence was measured to normalize transfectionefficiency. Reporter assays were performed 48 h post-transfec-tion using the Dual-luciferase assay-system (Promega), nor-malized for transfection efficiency by co-transfected Renillaluciferase.

DNA/RNA/protein isolation and cDNA synthesis from cells andfresh snap-frozen NSCLC specimensProtein isolation and western blot analysis were performed asdescribed by Mudduluru et al. (2010) using specific antibodiesof Axl (#sc-1096) or b-actin (#sc-1616-R) (sc: Santa CruzBiotechnology, Santa Cruz, CA, USA). DNA isolation andpurification from cells were performed with DNeasy Blood

4

R=0.72; p<0.001

miR-199b vs miR-199a

4

miR-199b vs miR-34a

5

R=0.34; p=0.02 R=0.65; p<0.001

a

e f

b c dmiR-34a vs miR-199a

p=0.02 p=0.01p=0.03

-2

0

2

-12 -2 8-2

0

2

-10 0 10-5

0

-12 -2 8

miR-199aS A AS

S

*

5

0

-5

-10

DD

CT

* *

-6

-4

miR-199b expression

miR

-199

a ex

pre

ssio

n

-6

-4

miR-199b expression

miR

-34a

exp

ress

ion

-10miR-34a expression

miR

-199

a ex

pre

ssio

n

miR-199bA

A: Adenocarcinoma (n=32)S: Squamous cell carcinoma (n=12)

miR-34a

Axl-upregulatedmiR-34ap = 0.07

100

8060

100

8060

100

8060

50403020100

50403020100

50403020100

p = Not Significant

miR-199a

Time

Per

cen

tsu

rviv

al

Per

cen

tsu

rviv

al

Per

cen

tsu

rviv

al

Time

miR-199bp = Not Significant

Down-regulated (≥ 1.5 fold )Up-regulated (< -1.5 fold)

Up-regulated (≥ 1.5 fold )

Down-regulated (< -1.5 fold)

Up-regulated (≥ 1.5 fold )Down-regulated (< -1.5 fold)

TotalMembrane

3

6

5

1

8 62.5

14.37Time

Axl-downregulated

miR-34a status Axl localization %Cytoplasm

Up-regulated

Down-regulated

40 X 40 X

Figure 9 Axl and miRs quantification in NSCLC tumors. MiR-34a, miR-199a and miR-199b expression in NSCLC tumor tissues.RUNB6 served as a normalization control. (a–c) Expression levels are given as DDCT (tumor/normal ratio). R and P-values arecalculated by Spearman’s rank correlation method. (d) Box plot of miRs (miR-34a, miR-199a and miR-199b) expression in thepatients’ tumor tissues according to histotypes. (e) Kaplan–Meier survival curves for the NSCLC patients divided by the status of miRsregulation. (f) Immunohistochemical quantification of Axl in the patients with miR-34a upregulated and downregulated tumors(n¼ 15). The box represents the number of cases with membrane or cytoplasmic staining. The percentage represents the proportion ofcases with cytoplasmic staining in each group. A full colour version of this figure is available at the Oncogene journal online.

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2897

Oncogene

and Tissue Kit (Qiagen). Total RNA was isolated from celllines and from lung specimens with Trizol reagent (Invitrogen)according to the manufacturer’s instructions. Expression ofmature miR-34a and miR-199a/b were determined by theTaqMan miRNA-assay (Applied Biosystems, Foster City, CA,USA), and normalized using the 2�DDCt method relative toU6-snRNA (RNU6B). Axl gene expression was determined asdescribed previously by Mudduluru et al. (2010). Primers areprovided in Supplementary Table S3.

Migration, invasion and CAM assayAssays were performed as described before by Mudduluruet al. (2010). In brief, Control-miR- or miR-34a- or miR-199a-transfected cells were placed either for CAM assay on theupper CAM of 10-day-old chicken embryos for tumor growthand in vivo metastasis analysis or for transwell chambers(Costar, Corning, NY, USA) and rest of the assays wereperformed as described by Mudduluru et al. (2010), respec-tively. Light microscopy pictures of transfected cells weretaken at � 40 magnification for wound healing assays.

5-aza-dC treatment of cells, bisulfite conversion of DNA andmethylation analysis50-Aza-dC treatment, bisulfite conversion and methylationanalysis (PCR was performed using HotStarTaq Plus DNAPolymerase (#203605) from Qiagen, Hilden, Germany) wereperformed as described by Mudduluru and Allgayer (2008).Axl and miRs expression were quantified in comparison withDMSO-treated samples. CpG islands upstream of the transcrip-tion start site or pri-miR start site were determined with theCpG island searcher (http://www.uscnorris.com/cpgislands2/cpg.aspx), and PCR primers were designed using the Methpri-mer software (http://www.urogene.org/methprimer) B1000bpupstream to the miRs or Axl transcription start site. Primersequences are provided in Supplementary Table S4.

ImmunohistochemistryExpression levels of Axl protein were detected by using goatanti-Axl antibody (sc-1096; Santa Cruz Biotechnology) at 4 1Covernight followed by rabbit anti-goat (biotinylated immuno-globulin G (IgG)) secondary antibody at room temperature for

30min. Immunoreactions were revealed by a biotin-freedextran-chain detection system (Envision, DakoCytomation,Glostrup, Denmark), and developed using diaminobenzidineas the chromogen. One section with isotype-non-specific IgGserved as negative controls for normal and tumor tissues.H1299 cell line served as positive controls. Semiquantitativescaling was used with slight modification as describedpreviously by Wu et al. (2002). Results were categorized intothree groups according to the percentage of positively stainedcells: score 1 (no staining); score 2, o10% of positive tumorcells; score 3, 410% of positive tumor cells. Additionally,tumors were divided according to the localization of Axlstaining (membrane or cytoplasm).

Statistical analysisTo test differential miR-34a, miR-199a and miR-199b expres-sion between tumors and corresponding normal lung tissues,the DDCt method was used, and miRs were consideredsignificantly overexpressed when DDCt values were 4±0.75.To test significant associations between expression levels andclinical-pathological variables, the Mann–Whitney U test andthe Fisher’s exact tests were applied. Univariate analysis ofsurvival was done with the method of Kaplan and Meier. In alltests, the statistical significance was set at P¼ 0.05.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

HA was supported by Alfried Krupp von Bohlen und HalbachFoundation, Essen, Hella-Buhler-Foundation, Heidelberg,Dr Ingrid zu Solms Foundation, Frankfurt/Main, HectorFoundation, Weinheim, Germany, FRONTIER ExcellenceInitiative of the University of Heidelberg, the BMBF, Bonn,Germany, and Walter Schulz Foundation, Munich, Germany.We thank Erika Hillerich and Laura Nelson for excellent helpand critical appraisal of the manuscript.

References

Bommer GT, Gerin I, Feng Y, Kaczorowski AJ, Kuick R, Love REet al. (2007). p53-mediated activation of miRNA34 candidatetumor-suppressor genes. Curr Biol 17: 1298–1307.

Budagian V, Bulanova E, Orinska Z, Duitman E, Brandt K, Ludwig Aet al. (2005). Soluble Axl is generated by ADAM10-dependentcleavage and associates with Gas6 in mouse serum.Mol Cell Biol 25:9324–9339.

Burchert A, Attar EC, McCloskey P, Fridell YW, Liu ET. (1998).Determinants for transformation induced by the Axl receptortyrosine kinase. Oncogene 16: 3177–3187.

Chang TC, Wentzel EA, Kent OA, Ramachandran K, Mullendore M,Lee KH et al. (2007). Transactivation of miR-34a by p53 broadlyinfluences gene expression and promotes apoptosis.Mol Cell 26: 745–752.

Chen R, Alvero AB, Silasi DA, Kelly MG, Fest S, Visintin I et al.(2008). Regulation of IKKbeta by miR-199a affects NF-kappaBactivity in ovarian cancer cells. Oncogene 27: 4712–4723.

Cheung HH, Lee TL, Davis AJ, Taft DH, Rennert OM, Chan WY.(2010). Genome-wide DNA methylation profiling reveals novelepigenetically regulated genes and non-coding RNAs in humantesticular cancer. Br J Cancer 102: 419–427.

Cho WC. (2007). OncomiRs: the discovery and progress of micro-RNAs in cancers. Mol Cancer 6: 60.

Doench JG, Sharp PA. (2004). Specificity of microRNA targetselection in translational repression. Genes Dev 18: 504–511.

Dumont N, Wilson MB, Crawford YG, Reynolds PA, SigaroudiniaM, Tlsty TD. (2008). Sustained induction of epithelial tomesenchymal transition activates DNA methylation of genessilenced in basal-like breast cancers. Proc Natl Acad Sci USA 105:14867–14872.

Erson AE, Petty EM. (2008). MicroRNAs in development and disease.Clin Genet 74: 296–306.

Esteller M. (2007). Cancer epigenomics: DNA methylomes andhistone-modification maps. Nat Rev Genet 8: 286–298.

Eulalio A, Huntzinger E, Izaurralde E. (2008). Getting to the root ofmiRNA-mediated gene silencing. Cell 132: 9–14.

Feinberg AP, Tycko B. (2004). The history of cancer epigenetics. Nat

Rev Cancer 4: 143–153.Fridell YW, Villa Jr J, Attar EC, Liu ET. (1998). GAS6 induces Axl-

mediated chemotaxis of vascular smooth muscle cells. J Biol Chem

273: 7123–7126.Gallardo E, Navarro A, Vinolas N, Marrades RM, Diaz T,

Gel B et al. (2009). miR-34a as a prognostic marker of relapse insurgically resected non-small-cell lung cancer. Carcinogenesis 30:1903–1909.

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2898

Oncogene

Garzia L, Andolfo I, Cusanelli E, Marino N, Petrosino G, De MartinoD et al. (2009). MicroRNA-199b-5p impairs cancer stem cellsthrough negative regulation of HES1 in medulloblastoma. PLoSOne 4: e4998.

Goruppi S, Ruaro E, Varnum B, Schneider C. (1997). Requirement ofphosphatidylinositol 3-kinase-dependent pathway and Src for Gas6-Axl mitogenic and survival activities in NIH 3T3 fibroblasts. Mol

Cell Biol 17: 4442–4453.Gupta GP, Massague J. (2006). Cancer metastasis: building a

framework. Cell 127: 679–695.Hafizi S, Dahlback B. (2006). Gas6 and protein S. Vitamin K-

dependent ligands for the Axl receptor tyrosine kinase subfamily.FEBS J 273: 5231–5244.

He L, He X, Lim LP, de Stanchina E, Xuan Z, Liang Y et al. (2007). AmicroRNA component of the p53 tumour suppressor network.Nature 447: 1130–1134.

Lee WP, Wen Y, Varnum B, Hung MC. (2002). Akt is required forAxl-Gas6 signaling to protect cells from E1A-mediated apoptosis.Oncogene 21: 329–336.

Li Y, Guessous F, Zhang Y, Dipierro C, Kefas B, Johnson E et al.(2009). MicroRNA-34a inhibits glioblastoma growth by targetingmultiple oncogenes. Cancer Res 69: 7569–7576.

Lodygin D, Tarasov V, Epanchintsev A, Berking C, Knyazeva T,Korner H et al. (2008). Inactivation of miR-34a by aberrant CpGmethylation in multiple types of cancer. Cell Cycle 7: 2591–2600.

Lujambio A, Calin GA, Villanueva A, Ropero S, Sanchez-CespedesM, Blanco D et al. (2008). A microRNA DNA methylationsignature for human cancer metastasis. Proc Natl Acad Sci USA105: 13556–13561.

Lujambio A, Esteller M. (2009). How epigenetics can explain humanmetastasis: a new role for microRNAs. Cell Cycle 8: 377–382.

McCloskey P, Fridell YW, Attar E, Villa J, Jin Y, Varnum B et al.(1997). GAS6 mediates adhesion of cells expressing the receptortyrosine kinase Axl. J Biol Chem 272: 23285–23291.

Melaragno MG, Cavet ME, Yan C, Tai LK, Jin ZG, Haendeler J et al.(2004). Gas6 inhibits apoptosis in vascular smooth muscle: role ofAxl kinase and Akt. J Mol Cell Cardiol 37: 881–887.

Mudduluru G, Allgayer H. (2008). The human receptor tyrosinekinase Axl gene—promoter characterization and regulation ofconstitutive expression by Sp1, Sp3 and CpG methylation. BiosciRep 28: 161–176.

Mudduluru G, Vajkoczy P, Allgayer H. (2010). Myeloid zinc finger 1induces migration, invasion, and in vivometastasis through Axl geneexpression in solid cancer. Mol Cancer Res 8: 159–169.

O’Bryan JP, Fridell YW, Koski R, Varnum B, Liu ET. (1995). Thetransforming receptor tyrosine kinase, Axl, is post-translationallyregulated by proteolytic cleavage. J Biol Chem 270: 551–557.

Raver-Shapira N, Marciano E, Meiri E, Spector Y, Rosenfeld N,Moskovits N et al. (2007). Transcriptional activation of

miR-34a contributes to p53-mediated apoptosis. Mol Cell 26:731–743.

Sainaghi PP, Castello L, Bergamasco L, Galletti M, Bellosta P, AvanziGC. (2005). Gas6 induces proliferation in prostate carcinoma celllines expressing the Axl receptor. J Cell Physiol 204: 36–44.

Stenhoff J, Dahlback B, Hafizi S. (2004). Vitamin K-dependent Gas6activates ERK kinase and stimulates growth of cardiac fibroblasts.Biochem Biophys Res Commun 319: 871–878.

Tazawa H, Tsuchiya N, Izumiya M, Nakagama H. (2007). Tumor-suppressive miR-34a induces senescence-like growth arrest throughmodulation of the E2F pathway in human colon cancer cells. ProcNatl Acad Sci USA 104: 15472–15477.

Tili E, Michaille JJ, Gandhi V, Plunkett W, Sampath D, Calin GA.(2007). miRNAs and their potential for use against cancer and otherdiseases. Future Oncol 3: 521–537.

Toyota M, Suzuki H, Sasaki Y, Maruyama R, Imai K, Shinomura Yet al. (2008). Epigenetic silencing of microRNA-34b/c and B-celltranslocation gene 4 is associated with CpG island methylation incolorectal cancer. Cancer Res 68: 4123–4132.

Vajkoczy P, Knyazev P, Kunkel A, Capelle HH, Behrndt S, vonTengg-Kobligk H et al. (2006). Dominant-negative inhibition of theAxl receptor tyrosine kinase suppresses brain tumor cell growthand invasion and prolongs survival. Proc Natl Acad Sci USA 103:5799–5804.

Valeri N, Vannini I, Fanini F, Calore F, Adair B, Fabbri M. (2009).Epigenetics, miRNAs, and human cancer: a new chapter in humangene regulation. Mamm Genome 20: 573–580.

van Ginkel PR, Gee RL, Shearer RL, Subramanian L, Walker TM,Albert DM et al. (2004). Expression of the receptor tyrosinekinase Axl promotes ocular melanoma cell survival. Cancer Res 64:128–134.

Weber M, Hellmann I, Stadler MB, Ramos L, Paabo S, Rebhan Met al. (2007). Distribution, silencing potential and evolutionaryimpact of promoter DNA methylation in the human genome. Nat

Genet 39: 457–466.Weinberg RA. (1995). The molecular basis of oncogenes and tumor

suppressor genes. Ann N Y Acad Sci 758: 331–338.Welch C, Chen Y, Stallings RL. (2007). MicroRNA-34a functions as a

potential tumor suppressor by inducing apoptosis in neuroblastomacells. Oncogene 26: 5017–5022.

Wu CW, Li AF, Chi CW, Lai CH, Huang CL, Lo SS et al. (2002).Clinical significance of AXL kinase family in gastric cancer.Anticancer Res 22: 1071–1078.

Yeligar S, Tsukamoto H, Kalra VK. (2009). Ethanol-inducedexpression of ET-1 and ET-BR in liver sinusoidal endothelial cellsand human endothelial cells involves hypoxia-inducible factor-1alpha and microrNA-199. J Immunol 183: 5232–5243.

Zhao Y, Srivastava D. (2007). A developmental view of microRNAfunction. Trends Biochem Sci 32: 189–197.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

miR-34a and 199a/b regulates Axl expressionG Mudduluru et al

2899

Oncogene